But some people believe there might still be some gold left in humanity’s pile of underused patents. Which is why computer scientists at Carnegie Mellon University and the Hebrew University of Jerusalem are working on a way to let computers mine old databases of patents and inventions to find ideas that can be repurposed to solve new problems.
The goal is to teach a computer system to find “analogies” – comparisons between different methods and problems that showcase their similarities. That’s not an easy task, even for humans. But by combining crowdsourcing and deep learning techniques, the team believe that they’ve made a breakthrough.
They devised a system where an army of volunteers hired through Amazon’s Mechanical Turk looked for products that had similar purposes or descriptions in Quirky.com’s product catalogue. Based on that data, they were able to teach a computer to do the same task – identifying analogies between seemingly-disparate products.
“After decades of attempts, this is the first time that anyone has gained traction computationally on the analogy problem at scale,” Aniket Kittur, associate professor in CMU’s Human-Computer Interaction Institute, who worked on the system.
They now hope that the same principle could be applied to patents and inventions. “Once you can search for analogies, you can really crank up the speed of innovation,” added Dafna Shahaf, a CMU alumnus and a computer scientist at Hebrew University, who also contributed to the research.
“If you can accelerate the rate of innovation, that solves a lot of other problems downstream.”